CONSOLE.

LIVE BRIGHT DATA RUN · NOT A REPLAY

EVIDENCE COVERAGE

100%

every published claim verified

FALSE REPORTS BLOCKED

incl. one overfit repair

GENUINE RUNS

4 published · 4 blocked

SELF-HEAL ROUNDS

2 human-approved in production

COLLECTOR

ROBUST

c_mt4m8fix1gze0scg44

SENSOR

ONLINE

live runs enabled

RUN 09j_competitor_moveTRUSTED CHANGEmonitoring

—

Different dataset, healthy sensor: HelioSupport overtook NimbusDesk with new benchmark evidence. Classified TRUSTED_SOURCE_CHANGE and published — Magpie does not heal real market movement. (Meta overwritten by later probe; rows preserved.)

fixture carousel · competitor_move · rev 26rows 10ranks 1–10failed none

RUN 08j_mt4s453e2aktj9wokyTRUSTEDverification

2026-08-22 19:36:25

Verification after heal round V4 (human-approved, 'both layouts'). Ten rows, hash-identical to baseline, robust across layouts.

fixture carousel · baseline · rev 25rows 10ranks 1–10failed none

RUN 07j_mt4rnmge14cpv5ma5gQUARANTINEDprobe

—

Regression probe on the original layout: the V3 repair expected the carousel and returned zero rows. The trust gate caught the overfit repair — quarantined, never published.

fixture cards · baseline · rev 24rows 0ranks —failed record_count

RUN 06j_mt4rk3d529os6d89ydTRUSTEDverification

—

Verification after heal round V3 (human-approved). All ten rows restored on the redesigned page; business-facts hash identical to baseline. Later found overfit to the carousel — see run 7.

fixture carousel · baseline · rev 22rows 10ranks 1–10failed none

RUN 05j_mt4qvbykzs0z36ag6QUARANTINEDmonitoring

2026-08-22 19:01:34

The believable lie. Transport-successful, schema-valid, seven complete rows — ranks 4–10. NimbusDesk silently missing. Quarantined; incident inc_001 opened.

fixture carousel · baseline · rev 22rows 7ranks 4–10failed record_count, rank_start, missing_ranks

RUN 04j_mt4qic011pisx5z7q6EMPTYmonitoring

—

Redesigned page iteration 2. Still zero rows — the parser keyed on a container id the redesign had dropped.

fixture carousel · baseline · rev 22rows 0ranks —failed record_count

RUN 03j_mt4mw9qqfffiyex5dEMPTYmonitoring

2026-08-22 18:51:28

First run against the redesigned page (iteration 1). Worker failed before extraction: zero rows, one input failure.

fixture carousel · baseline · rev 22rows 0ranks —failed record_count

RUN 02j_mt4mskyc7o888bkbaTRUSTEDbaseline

2026-08-22 17:07:27

Authoritative baseline. Ten vendors, ranks 1–10, full evidence. Operator-confirmed trusted reference.

fixture cards · baseline · rev 21rows 10ranks 1–10failed none

RUN 01j_mt4mgbxc1mq8zs3w4aDIAGNOSTICsmoke

—

First Collection API proof. Generation-era collector extracted a single schema-perfect row — led to the 'every card' regeneration.

fixture cards · baselinerows 1ranks 1–1failed record_count